Kinematics and Dynamics of the M 51-Type Galaxy Pair NGC 3893/96 (KPG 302)

A&A 466, 847–854 (2007) Astronomy DOI: 10.1051/0004-6361:20077071 & c ESO 2007 Astrophysics

Kinematics and dynamics of the M 51-type galaxy pair NGC 3893/96 (KPG 302)

I. Fuentes-Carrera1,, M. Rosado2,P.Amram3,H.Salo4, and E. Laurikainen4

1 Instituto de Astronomía, Geofísica e Ciencias Atmosféricas, Universidade de São Paulo, Rua do Matão 1226-Cidade Universitária, 05508-900 São Paulo SP, Brazil e-mail: [email protected], [email protected] 2 Instituto de Astronomía, Universidad Nacional Autónoma de México (UNAM), Apdo. Postal 70-264, 04510, México, D.F., México e-mail: [email protected] 3 Laboratoire d’Astrophysique de Marseille, 2 place Le Verrier, Marseille Cedex 4, France e-mail: [email protected] 4 Department of Physical Sciences, Division of Astronomy, University of Oulu, 90570 Oulu, Finland e-mail: [email protected],[email protected] Received 9 January 2007 / Accepted 23 January 2007

ABSTRACT

Aims. We study the kinematics and dynamics of the M 51-type interacting galaxy pair KPG 302 (NGC 3893/96). We analyze the perturbations induced by the encounter on each member of the pair, as well as the distribution of the dark matter (DM) halo of the main galaxy in order to explore possible differences between DM halos of “isolated” galaxies and those of galaxies belonging to a pair. Methods. The velocity field of each galaxy was obtained using scanning Fabry-Perot interferometry. A two-dimensional kinematic and dynamical analysis of each galaxy and the pair as a whole are done emphasizing the contribution of circular and non-circular velocities. Non-circular motions can be traced on the rotation curves of each galaxy allowing us to differentiate between motions associated to particular features and motions that reflect the global mass distribution of the galaxy. For the main galaxy of the pair, NGC 3893, optical kinematic information is complemented with HI observations from the literature to build a multi-wavelength rotation curve. We try to fit this curve with a mass-distribution model using different DM halos. Results. Non-circular motions are detected on the velocity fields of both galaxies. These motions can be associated to perturbations due to the encounter and, in the case of the main galaxy, to the presence of a structure such as spiral arms. The location of the corotation radius of this galaxy is also explored. We find that the multi-wavelength rotation curve of NGC 3893, “cleaned” from the effect of non-circular motions, cannot be fitted whether by a pseudo-isothermal or by a NFW DM halo. Key words. galaxies: interactions – galaxies: kinematics and dynamics – galaxies: individual: NGC 3893, NGC 3896 – galaxies: spiral – galaxies: halos

1. Introduction the classical method for studying mass distribution, see Blais- ff Ouellette et al. (2001), and references therein. The RCs also al- The di erence between the mass distribution implied by the lu- low us to determine the maximum rotation velocity of a galaxy minosity of a disk galaxy and the distribution of mass implied ff and thus infer the total mass within a certain radius using meth- by the rotation velocities o ers strong evidence that disk galax- ods such as that of Lequeux (1983). Nevertheless, care must be ies are embedded in extended halos of dark matter (Sofue & taken when using kinematic information from interacting galax- Rubin 2001, and references therein). Detailed knowledge of dark ies, since they are subject to kinematical perturbations that may matter (DM) halos around galaxies holds important clues to the affect the correct determination of an RC that actually traces the physics of galaxy formation and evolution and is an essential in- global mass distribution of the galaxy. For this reason, 3D spec- gredient for any model aiming to link the observable Universe troscopy observations are required to separate circular from non- with cosmological theories. In practice, realistic DM halos are circular motions in the velocity field of a galaxy and its rotation neither static nor spherically symmetric (Knebe et al. 2004) and curve as shown in Fuentes-Carrera et al. (2004). it is still unknown if their structure and distribution is intrinsi- cally related to the environment of their galaxies. The question In this work, we present scanning Fabry-Perot observa- / remains as to whether an intrinsic difference exists between the tions of the M 51-type interacting galaxy pair NGC 3893 96 DM halo of an “isolated” galaxy, the DM halo of a galaxy be- (KPG 302). Section 2 presents the scanning Fabry-Perot (FP) longing to a pair or that of a galaxy that is part of a larger group observations and data reductions. Section 3 introduces the pair / such as a compact group or a cluster. of galaxies KPG 302 (NGC 3893 96). In Sect. 4 we present In this sense, rotation curves (RCs) are a powerful tool the kinematic information derived from the F-P observations. for studying the distribution of matter (both baryonic and non- Section 5 presents the dynamical analysis of both galaxies, mass baryonic) in interacting disk galaxies. For a description of estimates and the mass distribution for NGC 3893. The discussion and conclusions are presented in Sects. 6 and 7, respec- Presently at the Observatoire de Paris-Meudon. tively.

Article published by EDP Sciences and available at http://www.aanda.org or http://dx.doi.org/10.1051/0004-6361:20077071 848 I. Fuentes-Carrera et al.: Galaxy pair KPG 302

2. Observations and data reductions Observations of NGC 3893/96 (KPG 302) were done at the 2.1 m telescope at the OAN-SPM (México) using the scanning Fabry-Perot interferometer PUMA (Rosado et al. 1995). PUMA is a focal reducer built at the Instituto de Astronomía-UNAM used to make direct images and Fabry-Perot (FP) interferometry of extended emission sources (field of view 10). The FP used is an ET-50 (Queensgate Instruments) with a servostabilization system having a free spectral range of 19.95 Å (912 km s−1)at Hα. Its finesse (∼24) leads to a sampling spectral resolution of 0.41 Å (19.0 km s−1) achieved by scanning the interferometer free spectral range through 48 different channels. A 1024 × 1024 Tektronix CCD detector with a resolution of 0.58/pixel was used. We used a 2 × 2 binning to enhance the signal. The fi- nal spatial sampling equals 1.16/pixel. To isolate the redshifted Hα (λat rest = 6562.73 Å) emission of the galaxies, we used an interference filter centered at 6584 Å with an FWHM of 10 Å. / To average the sky variations during the exposure, we got two Fig. 1. a) Direct B image of NGC 3893 96 (KPG 302) from “The data cubes with an exposure time of 48 min each (60 s per chan- Carnegie Atlas of Galaxies. Volume II” (Sandage & Bedke 1994). b) Monochromatic Hα (continuum subtracted) image of the pair ob- nel). These data cubes were co-added leading to a total exposure tained from the scanning Fabry-Perot interferometer PUMA data cubes. time of 96 minutes. For the calibration, we used a H lamp whose Upper panel: Optical image with HI isophotes superposed. Image 6562.78 Å line was close to the redshifted nebular wavelength. taken from Verheijen & Sancisi (2001) in “An HI Rogues Gallery” Two calibration cubes were obtained at the beginning and at the (http://www.nrao.edu/astrores/HIrogues/webGallery/ end of the galaxy observation to check the metrology. RoguesGallery06.html) Data reduction and analysis were done using mainly the ADHOCw1 and CIGALE softwares (LeCoarer et al. 1993). particular concentration toward the center. Given the isolation Standard corrections (cosmic rays removal, bias subtraction, criteria used in the KPG and the nature of the cluster, it is pos- flat-fielding, etc.) were done on each cube. Once the object sible that the DM halo (or halos) of the galaxies in the pair are cubes were co-added, the night sky continuum and 6577.3 Å isolated from those of the cluster. OH sky line were subtracted. A spectral Gaussian smoothing = −1 NGC 3893 is a grand-design spiral similar to NGC 5194 (σ 57 km s ) was also performed. Once the spectral smooth- in M 51 (Fig. 1a). It has been classified as SABc in LEDA ing was done, the calibration in wavelength was fixed for each database, and as SAB(rs)c in the NED2 database and in the profile at each pixel using the calibration cube. The Fabry-Perot RC3 (de Vaucouleurs et al. 1991). However, Hernández-Toledo scanning process allows us to obtain a flux value at pixel level for & Puerari (2001) classify it as a non-barred galaxy without any each of the 48 scanning steps. The intensity profile found along inner ring. The HI observations (Verheijen & Sancisi 2001) show the scanning process contains information about the monochro- that this galaxy is slightly warped in its outer parts – both in its matic emission (Hα) and the continuum emission of the object. HI distribution and its HI kinematics. Its companion, NGC 3896, The continuum image computation was done considering the appears to be an intermediate type galaxy between and S0 and mean of the 3 lowest intensities of the 48 channels cube. For a spiral, which also has a bar. It shows extended Hα emission. the monochromatic image, the Hα line intensity was obtained NGC 3893 shows no color excess, whereas NGC 3896 has a by integrating the monochromatic profile in each pixel. The ve- predominantly blue B − V color in the central parts of the galaxy locity maps were computed using the barycenter of the Hα pro- ffi (Laurikainen et al. 1998; Hernández-Toledo & Puerari 2001). file peaks at each pixel. To get a su cient signal-to-noise ratio The star-formation rate (SFR) of each member of the pair on the outer parts of each galaxy, we performed three spatial was derived by James et al. (2004). These authors find a value of Gaussian smoothings (σ = 2.36, 3.54, 4.72 ) on the resulting −1 −1 5.62 M yr for NGC 3893 and a value of 0.14 M yr for its calibrated cube. A variable-resolution radial velocity map was companion. NGC 3893 was also part of a dynamical analysis of built using high spatial resolution (less spatially-smoothed pix- high surface brightness spiral galaxies using long-slit observa- els) for regions with an originally higher signal-to-noise ratio. tions and numerical modeling in order to quantify the luminous- to-dark matter ratio inside their optical radii (Kranz et al. 2003). 3. NGC 3893 and NGC 3896 According to these authors, NGC 3893 has a massive stellar disk that dominates the dynamics of the central regions with a disk NGC 3893/96 is an interacting galaxy pair with number 302 mass of 2.32 × 1010 solar masses. They infer the location of the in the Catalogue of Isolated Pairs of Galaxies in the Northern corotation resonance of this galaxy at 5.5 ± 0.5 kpc. Though op- Hemisphere (KPG, Karachentsev 1972). Morphologically, it re- tical images of this pair do not show an apparent connection be- sembles M 51 (NGC 5194/95) since it is composed of a main tween the two galaxies, radio images by Verheijen & Sancisi spiral galaxy and a considerably smaller companion. KPG 302 (2001) show extended HI emission encompassing both galaxies is situated in the Ursa Major cluster. With only 79 members, (small panel in Fig. 1). This common HI envelope is elongated this cluster is poorly defined with a velocity dispersion of only from SE to NW, parallel to the line that joins the nuclei of both − 148 km s 1 and a virial radius of 880 kpc (Tully et al. 1996). It galaxies. HI isophotes also show what could be considered as a essentially contains only late-type galaxies distributed with no 2 The NASA/IPAC Extragalactic Database (NED) is operated by the 1 http://www.oamp.fr/adhoc/adhocw.htm developed by J. Jet Propulsion Laboratory, California Institute of Technology, under Boulesteix. contract with the National Aeronautics and Space Administration. I. Fuentes-Carrera et al.: Galaxy pair KPG 302 849

Table 1. Parameters of NGC 3893 and NGC 3896.

NGC 3893 NGC 3896 Coordinates (J2000)a α = 11h 48m 38.38 α = 11h 48m 56.42s δ =+48◦4234.4 δ =+48◦4029.2 Morphological type SABca S0-aa SAB(rs)cb,c S0/ac:pecb Scd SBbc pecd d mB (mag) 11.13 14.05 B − Vd (in mag) 0.56 0.46 a D25/2 ( )2.03 0.75 Distancee (Mpc) 18.618.6 f SFR (M/yr) 5.62 0.14 Heliocentric systemic velocity (km s−1) 969 ± 3a,g 959 ± 12a,g 973j --- 962.0 ± 5h 920 ± 5h 958.0k --- −1 a,g a,g Vrotmax (km s ) 251.2 150 195j con f usedj 220 ± 3l --- 197 ± 10h 50 ± 10h 207k --- PA ( ◦) 352j --- 166 + 180m --- 340 ± 10h 294 ± 5h 167 + 180k --- Inclination (◦)49± 2i 48 ± 3i 42m --- 45 ± 3h 49 ± 3h 49k --- n mK (mag) 7.891 11.648 a LEDA database. b NED database. c RC3 (de Vaucouleurs et al. (1991). d Hernández-Toledo & Puerari (2001). e Tully & Pierce (2000). f James et al. (2004), total measured Hα + [NII] line flux corrected for [NII] contamination. g From HI observations. h This work. i Verheijen (2001), HI observations. j Verheijen & Sancisi (2001), HI observations. k Garrido et al. (2002). l Tully et al. (1996). m Kranz, Slyz & Rix (2003). n 2MASS (Skrutskie et al. 2006) Kext value taken from the NED database. broad arm going from NGC 3893 to NGC 3896. Table 1 lists the axis is almost perpendicular to the main axis of the galaxy, out- main parameters of each galaxy. lined by isovelocities with 955 km s−1. Small wiggles seen along these isovelocities could be a signature of an inner ring, similar to what is shown by the simulations by Salo et al. (1999) 4. Kinematic results of IC4214. The velocity field is symmetrical with respect to the 4.1. Monochromatic images kinematical minor axis. Locally, minor irregularities are seen in the distribution of radial velocities, especially along the spiral Figure 1b displays the monochromatic Hα image of the pair. arms of the galaxy. These might be associated to the passage of Knotty HII regions lie along the spiral arms of NGC 3893. gas through the spiral density wave. For NGC 3896, the velocity Although they mainly follow the two main arms, they also out- field is very perturbed, displaying a mild velocity gradient from line small segments of fainter and less pronounced flocculent the SE to the NW (upper panel in Fig. 3). Isovelocities are very arms. Intense HII regions are seen on the east side of the galaxy patchy and crooked. along the western spiral arm. A very bright HII region is seen in the central parts of the galaxy. For NGC 3896, significant Hα emission is seen within the 4.3. Rotation curves inner 13.9 (1.3 kpc) southeast of the center of the galaxy. This emission displays two maxima that form a rather elongated re- In the case of early-stage interactions, the inner parts of galax- gion (left panel in Fig. 1b). The northern parts of the galaxy show ies are not strongly perturbed, velocity fields are still smooth weak diffuse emission. This type of emission is also seen on the and symmetrical, resulting in symmetric and low-scattered rota- western side of the galaxy. These sides are closer to NGC 3893. tion curves (RCs) up to a certain radius (Fuentes-Carrera et al. 2004). With this assumption in mind, the RC of each galaxy was ff 4.2. Velocity fields computed considering di erent values for the kinematical parameters involved in order to obtain a symmetric curve in the The upper panel in Fig. 2 shows the velocity field of NGC 3893. inner parts of the galaxy and to minimize scatter on each side of It is a smooth and regular field; the isovelocities show no sig- the curve. The rotation curves are sampled with bins of 2 pixels nificant distortions. Although the galaxy has an elongated inner (∼2.3). Error bars give the dispersion of the rotation velocities structure in the monochromatic image (see Fig. 1), the central computed for all the pixels found inside each elliptical ring de- parts of the galaxy show no kinematic signatures that could be fined by the successive bins. This approach is described in more associated to the presence of a nuclear bar. The minor kinematic detail in Fuentes-Carrera et al. (2004) and Garrido et al. (2005). 850 I. Fuentes-Carrera et al.: Galaxy pair KPG 302

Fig. 2. Top: velocity field of NGC 3893 in KPG 302. The solid line Fig. 3. Top: velocity field of NGC 3896 in KPG302. The solid line in- indicates the galaxy’s position angle (PA) and the slash-dotted lines indicates the galaxy’s position angle (PA) and slash-dotted lines indicate dicate the angular sectors from both sides of the major axis considered the angular sectors from both sides of the major axis considered for the for the computation of the galaxy’s the computation of the galaxy’s ro- computation of the galaxy’s RC. Bottom: RC of NGC 3893. Both sides tation curve. Bottom: rotation curve (RC) of NGC 3893. Both sides of of the curve have been superposed. Open squares correspond to the re- the curve have been superposed. Open squares correspond to the receding side of the galaxy. Filled squares correspond to the approaching ceding side of the galaxy. Filled squares correspond to the approaching ◦ ◦ side. RC was plotted considering an inclination value of 49 . The solid side. RC was plotted considering an inclination value of 45 . Horizontal arrow indicates maximal rotation velocity. solid arrow indicates the maximal rotation velocity. Vertical dotted arrow indicates the radius associated with corotation according to Kranz et al. (2003). 4.3.2. NGC 3896 Though the velocity field of this galaxy is very perturbed, we were still able to derive an RC reflecting the circular motions of 4.3.1. NGC 3893 the southwestern side of the galaxy (approaching side). The fol- = ± ◦ = ± ◦ The RC of NGC 3893 was computed considering points on the lowing set of values were used: PA (294 5) , i (49 3) , ◦ −1 and Vsys = (920.50 ± 5) km s , considering an angular sector velocity field within an angular sector of 32 on each side of ◦ the galaxy’s PA. The kinematic center used to compute the rota- of 42 on each side of the galaxy’s PA. These values, along with tion curve matches the photometric center from the PUMA con- values from previous works, are presented in Table 1. The lower tinuum image within 1. The most symmetrical, smooth, and panel in Fig. 3 shows the RC derived for this galaxy superposing both the approaching and receding sides. Globally the curve less-scattered RC was derived using the following set of values: ◦ ◦ −1 increases up to the last emission point at 17.5 (1.5 kpc) where PA = (340±10) , i = (45±3) and Vsys = (962±5) km s .These the velocity equals 48 km s−1. Rather low velocity values appear are presented in Table 1 along with values obtained in previous works. The RC superposing both the approaching and receding near the center of the galaxy, within 4.2 (0.4 kpc). sides is shown in the lower panel in Fig. 2. Globally the RC of NGC 3893 is symmetric up to the last Hα emission point at 85 4.4. Non-circular motions (7.55 kpc). Both sides of the curve display oscillations of about 10 km s−1. The maximum rotation velocity is (197 ± 10) km s−1 Two-dimensional kinematic fields of disk galaxies portray the and is reached at 69 (6.1 kpc). Our RC is very similar to the one motion of the gas over the whole galaxy enabling us to match derived by Garrido et al. (2002) using FP observations. To com- these motions with different morphological structures. One can pare with the results derived by Kranz et al. (2003), a long slit determine to what extent the gas is following circular motion was simulated by considering radial velocities on the velocity around the center of the galaxy and to what extent there are large field within a sector of 1.5◦ on each side of the slit position an- contributions from non-circular velocities (radial, azimuthal, and gle. The values of the kinematic parameters were equal to those vertical ) due to the presence of these structures or to external used by these authors. These “long-slit” RCs show the same in- perturbations. For NGC 3893, we analyzed the influence of mor- creasing behavior and oscillations as those found by Kranz et al. phological features on the kinematics of the gas. This was done (2003). by comparing the monochromatic image with each side of the I. Fuentes-Carrera et al.: Galaxy pair KPG 302 851

Fig. 5. Multi-wavelength curve of NGC 3893. Small dots in the inner parts of the curve correspond to optical Fabry-Perot Hα observations. Larger dots in the outer parts correspond to the HI curve derived by Verheijen & Sancisi (2001). The horizontal arrow indicates the point with maximum rotation velocity. The vertical arrow shows the radius considered for mass estimation using the method by Lequeux (1983).

distribution). As a first estimate, we considered the maximum rotation velocity (190 km s−1) given by both the optical and ra- Fig. 4. Top: Monochromatic image of NGC 3893. Letters indicate features associated to points on the RC (shown in the bottom panel)in dio RCs to estimate the mass within R = 120 = 10.8 kpc = order to differentiate the contribution of circular from non-circular mo- 0.99 D25/2. The range of masses within this radius is given by 11 tions. Solid line indicates the galaxy’s position angle (PA) and the thin 0.50 to 0.84 × 10 M. To derive the mass of NGC 3896, we lines indicate the angular sector from both sides of the major axis con- considered the last emission point on the approaching side of the sidered for the computation of the galaxy’s RC. curve, which shows the maximum amplitude of that curve. This −1 value equals 48 km s at R = 17.5 = 1.5 kpc = 0.4 D25/2. The × 8 RC (Fig. 4). Through this comparison we were able to differen- masses within this radius range from 4.78 to 7.97 10 M. tiate points in the RC associated to circular motions of the gas – The mass ratio of the galaxies was computed from the associated with the global mass distribution of the galaxy – from NIR luminosities, calculated from the Ks-magnitudes taken from 2MASS (Skrutskie et al. 2006). These luminosities for non-circular motions – associated to the response of the gas to / local morphological features. NGC 3893 96 lead to a mass ratio of 0.031. On the other hand, the mass ratio derived from the RC of each galaxy within 0.4 D25/2 falls around 0.0255. These values are fairly similar, 5. Dynamical analysis particularly if considering the uncertainty of the mass-to-light ratio. This strengthens the reliability of the mass estimates based 5.1. Mass estimates through rotation curves on the rotation curves. The Hα kinematic information from our FP observations of NGC 3893 was complemented with HI synthesis observations 5.2. Mass distribution by Verheijen & Sancisi (2001). To match both curves, we considered the averaged optical rotation curve that was derived us- We used the mass model from Blais-Ouellette et al. (2001) to ing the same kinematic parameters as those used for the HI RC study the mass distribution in NGC 3893. This model uses both (these values do not differ much from the values we used for the the light distribution of the galaxy and a theoretical dark-halo optical RC – see Table 1). The superposed curves are shown in profile to compute an RC that best-fits the observed one. The Fig. 5. Both curves superpose smoothly except for the external mass-to-light ratio of the disk (M/L)disk), as well as the prop- parts of the optical RC (R ≥ 60 ), which appear to be more per- erties of the dark matter, characteristic density (ρ0), and radius turbed. The HI RC also decreases for R ≥ 175 anduptothelast (R0), are free parameters. We used a DM halo described by a HI emission point at R = 245 (21.76 kpc). Verheijen & Sancisi pseudo-isothermal sphere (Begeman 1987) and a Navarro, Frenk (2001) mention that the determination of the HI rotation curve & White (NFW) profile (Navarro et al. 1996) to fit the multi- beyond 120 is uncertain because of the tidal interaction with wavelength RC of NGC 3893. Optical photometry in the I band NGC 3896. We set this radius as our limit for the computation was taken from Hernández-Toledo & Puerari (2001) and the HI of the mass through the RC of NGC 3893. superficial distribution from Verheijen & Sancisi (2001). Making Using this composite rotation curve, a range of possible use of the possibility of disentangling circular from non-circular masses was computed using the method proposed by Lequeux velocities in the optical RC, our multi-wavelength RC was (1983) according to which the total mass of a galaxy within a “cleaned” from points associated to non-circular motions (see radius R lies between 0.6 (in the case of a disk-like mass distri- Sect. 4.4) and to the warp in the outer parts of the HI disk. We bution) and 1.0 × (RV2(R)/G) (in the case of a spheroidal mass removed points in the optical part of the observed RC between 852 I. Fuentes-Carrera et al.: Galaxy pair KPG 302

Fig. 6. Best mass model ﬁts for the multi-wavelength rotation curve of NGC 3893 once the points associated with non-circular motions have been removed. Top left: pseudo-isothermal halo and non-maximal disk. Top right: pseudo-isothermal halo and maximal disk. Bottom left:NFWhalo and non-maximal disk. Bottom right: NFW halo and maximal disk. The long-dashed curve represents the dark-matter halo contribution and the short-dashed curve represents the stellar disk contribution. The parameters displayed stand for the mass-to-light ratio of the stellar disk (M/Ldisk), 2 the characteristic radius of the dark-matter halo and density (R0 and ρ0, respectively) and the minimized χ in the three-dimensional parameter space. Mass-model taken from Blais-Ouellette et al. (2001).

15 –20 ,22 –32 ,andatR > 75 (see Fig. 5). Only points ((M/L)disk = 0.56 – in both K and I bands). We must take into from the HI rotation curve were taken into account after 75 and account that Kranz et al. (2003) only fit the optical part of the up to 120 considering the fact that the RC in HI is uncertain RC derived with long-slit spectroscopy. For the sake of compar- beyond this radius. ison, we fitted our Hα RC (with and without points associated Figure 6 shows different fits for this multi-wavelength RC: a to non-circular motions) using the value for (M/L)disk derived pseudo-isothermal DM halo with non-maximal disk (top left), by these authors. Results are shown in Fig. 8 and Table 2. Fits a pseudo-isothermal DM halo with a maximal disk (top right), are very good for both the pseudo-isothermal and the NFW halo; = a NFW halo with non-maximal disk (bottom left), and an NFW nevertheless, (M/L)disk 0.56 does not render the disk maximal. halo with a maximal disk (bottom left). Table 2 displays the Finally we also fitted these mass models to the HI RC without ff ff mass-model parameters used in each case. We used the defini- points at R > 120 where tidal e ects might a ect the correct tion of Sackett (1997) for the “maximal disk" that is taken to be determination of the RC. Figure 9 and Table 2 show the result- a galactic disk such that 85% ± 10% of the total rotational sup- ing parameters of the fits. The fit using the NFW halo misses the 2 = port of a galaxy at a radius 2.2 × scale radius is contributed by innermost point of the curve and is not very accurate (χ 3.17). the stellar-disk mass component. For NGC 3893, this radius cor- The fit using the pseudo-isothermal halo misses the middle point 2 = responds to 1.80 kpc (Kranz et al. 2003). The best fit (χ2 = 1.34) of the curve, yet χ 1.32 – which is one of the lowest values is obtained using a pseudo-isothermal halo with a non-maximal found for all fits presented. Nevertheless it should be noticed that for both halos, the (M/L) > 1.7, which is higher than disk leading to (M/L)disk = 0.94 in the I band, yet it misses disk the last two points of the RC. The fit with an NFW halo and the value found with the multi-wavelength RC. This highlights 2 the importance of the multi-wavelength approach for the mass a non-maximal disk gives (M/L)disk = 0.24 (χ = 1.43), also missing the two outermost points of the RC. Both a pseudo- model. isothermal and NFW halo with a maximal disk give higher val- 2 ues of (M/L)disk (1.25 in both cases) and also higher χ values (1.53 and 3.11, respectively). They also miss both outer and in- 6. Discussion ner points on the RC. That no model fits the last point in the multi-wavelength RC of To evaluate the effect of non-circular motions on the fits, NGC 3893 “cleaned” from the effects of non-circular motions we fitted the above mass models by including the points asso- can be explained either by a truncated halo for this galaxy or ciated to non-circular motions to the multi-wavelength RC. The by the existence of a common halo for both galaxies. Since the fits can be seen in Fig. 7. The values for the mass-model pa- derived mass of the NGC 3896 is low, then it most probably re- rameters are shown in Table 2. These fits are less precise. In all sides inside the halo of NGC 3893, and both galaxies share a sin- cases, the mass-to-luminosity ratio using the maximal disk as- gle halo. Nevertheless this halo would have a different distribu- sumption is higher than the one obtained by Kranz et al. (2003) tion than in an isolated galaxy. When considering an isothermal I. Fuentes-Carrera et al.: Galaxy pair KPG 302 853

Table 2. Mass models parameters for NGC 3893 from best ﬁts of the multi-wavelength rotation curve, considering only points associated to circular motions.

2 Type of Maximal (M/L)disk R0 ρ0 χ −3 halo disk (I band) (kpc) (M/pc ) Multi-wavelength curve, non-circular motions excluded p-ISO no 0.940 1.570 0.340 1.34 p-ISO yes 1.250 2.000 0.210 1.53 NFW no 0.240 7.160 0.075 1.43 NFW yes 1.250 7.100 0.050 2.60 Multi-wavelength curve, non-circular motions included p-ISO no 1.050 1.100 0.710 2.16 p-ISO yes 1.250 1.500 0.310 1.83 NFW no 0.070 4.650 0.160 1.72 NFW yes 1.250 7.100 0.050 2.60 Hα curve, non-circular motions included p-ISO no 0.560 1.090 0.900 1.09 NFW no 0.560 10.100 0.040 0.96 HI curve, non-circular motions excluded p-ISO yes 1.750 2.600 0.079 1.32 NFW yes 1.745 7.000 0.030 3.17

Fig. 7. Best mass-model ﬁt for the multi-wavelength rotation curve of NGC 3893 considering all observed points – including those associated with non-circular motions. Top left: pseudo-isothermal halo and non-maximal disk. Top right: pseudo-isothermal halo and maximal disk. Bottom left: NFW halo and non-maximal disk. Bottom right: NFW halo and maximal disk.

halo and the HI curve, the (M/L)disk that ﬁts this curve best Hubble types of normal galaxy. This is the case for NGC 3896, is much higher than the M/L found for the multi-wavelength whose RC displays almost solid body rotation up to the last curve. In general, the information on the inner parts of the galaxy emission point. given by the optical observations imposes an (M/L)disk ≤ 1, which would imply the presence of a disk with a large pop- ulation of young stars -which is not the case given the B − V 7. Conclusions value derived by Hernández-Toledo & Puerari (2001). This sup- We have presented the kinematic and dynamical analysis of the ports the idea that the structure of the DM halo of this pair dif- M 51-type galaxy pair, KPG 302 (NGC 3893/96). NGC 3893 is fers from that of a single-disk galaxy. As shown by Laurikainen a grand-design spiral with a regular velocity ﬁeld that displays & Salo (2001), M 51-type pairs companions tend to have ex- no major distortions. The companion, NGC 3896, displays on- tremely large bulge sizes relative to their disk scale-lengths. going star formation that was probably triggered by the interac- Consequently, the bulge-to-disk luminosity ratios for the com- tion with the main galaxy. This galaxy displays important non- panions were also generally higher than known for any of the circular motions in localized regions, especially on the side of 854 I. Fuentes-Carrera et al.: Galaxy pair KPG 302

Acknowledgements. We wish to thank the staff of the Observatorio Astronómico Naconal (OAN-SPM) for their support during PUMA data acquisition. We also thank C. Carignan for letting us use his mass model. I.F.-C. acknowledges the financial support of FAPESP and CONACYT grants No. 03/01625-2 and No. 121551 – respectively. M.R. acknowledges financial support from grants 46054-F from CONACYT and IN100606 from DGAPA-UNAM. H.S. and E.L. acknowledge the support from the Academy of Finland. We acknowledge the use of the HyperLeda database (http://leda.univ-lyon1.fr), the NASA/IPAC Extragalactic Database (NED), and the Two Micron All Sky Survey.

Fig. 8. Best mass-model fit for the Hα rotation curve of NGC 3893 References considering all observed points -including those associated with non- Begeman, K. G. 1987, Ph.D. Thesis, Groningen University circular motions and using the (MŁ)disk value by Kranz et al. (2003). Left: pseudo-isothermal halo. Right:NFWhalo. Blais-Ouellette, S., Amram, P., & Carignan, C. 2001, AJ, 121, 1952 Fuentes-Carrera, I., Rosado, M., Amram, P., et al. 2004, A&A, 415, 451 Garrido, O., Marcelin, M., Amram, P., & Boulesteix, J. 2002, A&A, 387, 812 Garrido, O., Marcelin, M., Amram, P., Balkowski, C., & Boulesteix, J. 2005, MNRAS, 362, 127 Hernández-Toledo, H., & Puerari, I. 2001, A&A, 379, 54 James, P. A., Shane, N. S., Beckman, J. E., et al. 2004, A&A, 414, 23 Karachentsev, I. D. 1972, Catalogue of isolated pairs of galaxies in the northern hemisphere, Soobshch. Spets. Astrofiz. Obs, 7, 1 Knebe, A., Gill, S. P. D., & Gibson, B. K. 2004, PASA, 21, 216 Kranz, T., Slyz, A., & Rix, H.-W. 2003, ApJ, 586, 143 Laurikainen, E., & Salo, H. 2001, MNRAS, 324, 685 Laurikainen, E., Salo, H., & Aparicio, A. 1998, A&A, 129, 517 Le Coarer, E., Rosado, M., Georgelin, Y., Viale, A., & Goldes, G. 1993, A&A, Fig. 9. Best mass-model fit for the HI rotation curve of NGC 3893 after 280, 365 Lequeux, J. 1983, A&A, 125, 394 removing points associated with non-circular motions and the warp of Navarro, J. F., Frenk, C. S., & White, S. D. M. 1996, ApJ, 462, 563 the outer parts of the disk. Left: pseudo-isothermal halo. Right:NFW Pasha, I. I., & Smirnov, M. A. 1982, Ap&SS, 86, 215 halo. Rosado, M., Langarica, R., Bernal, A., et al. 1995, RevMexAA (Serie de Conferencias), 03, 263 Sackett, P. D. 1997, ApJ, 483, 103 the galaxy that is closer to the companion. The total mass of each Salo, H., Rautiainen, P., Buta, R., et al. 1999, AJ, 117, 792 galaxy was derived from the RC. The optical RC of NGC 3893 Sandage, A., & Bedke, J. 1994, The Carnegie Atlas of Galaxies. Volume II, was matched with the existing HI curve to determine the distri- Carnegie Institution of Washington with The Flintridge Foundation Skrutsie, M. F., Cutri, R. M., Stiening, R., et al. 2006, AJ, 131, 1163 bution of the luminous and dark matter. This multi-wavelength Sofue, Y., & Rubin, V. 2001, ARA&A, 39, 137 rotation curve was analyzed in light of the 3D observations in Struck, C. 2005 [arXiv:astro-ph/0511335] order to differentiate the contribution of non-circular motions Tully, R. B., Verheijen, M. A. W., Pierce, M. J., Huang, J.-S., & Wainscoat, R. J. associated to particular features and the contribution of circu- 1996, AJ, 112, 2471 lar motions, which reflect the mass distribution of the galaxy. de Vaucouleurs G., de Vaucouleurs A., Corwin H. G., et al. 1991, Third Reference Catalogue of Bright Galaxies (RC3) (New York: Springer-Verlag) No “classical” DM halo fits the observed rotation curve, which Verheijen, M. A. W. 2001, ApJ, 563, 694 could imply a different mass distribution for the DM halo of Verheijen, M. A. W., & Sancisi, R. 2001, A&A, 370, 765 M 51-type binary galaxies.